Dylan Schiemann

We use Daylite for our internal group calendar, and for the most part we love this Mac-only application. However, there are a few basic things about it that do not make sense.

• The fields for start and end time are not wide enough to see the start and end time (see image). Regardless of how large you set the dialog, there is no resizing mechanism for these fields. Instead, to get the information you need, you must click a schedule button that pulls up a dialog within a dialog.
• The default start and end time are set to the current clock time. By default, if you create a new event 14 days in the future, but your current clock time is 10:32, the start time it set to 10:32 and the end time one hour later.
• There are dialogs within dialogs within dialogs. In order to make the app feel simple and easy to use, significant features are hidden away in a number of dialogs. The problem is that a number of the icons for selecting dialogs do not make a lot of sense unless you schedule events with a significant amount of frequency. This also slows down the process of scheduling.
• The Daylite server can only be hosted on a Mac unless you purchase of a rather expensive DB license. We host it on a Mac mini, but we would rather server it through our normal linux servers, as I imagine most businesses would prefer.

Despite my complaints, we use Daylite extensively for scheduling and contact management, and enjoy using it. It has a number of other features (projects, opportunities, tasks) that we don’t use because we find that those features are better solved through issue tracking systems that integrate with subversion. Some may find it interesting that we prefer a desktop application for this feature set. We’ve basically found that none of the web-based apps out there match the core feature set we need, and provide the level of privacy we receive by hosting the service ourselves.

Thunderbird 2.0 RC1 has significant performance and interface improvements, including a filtering mechanism for displaying message folders. I find myself in a view that shows only folders with unread messages 90+% of the time. A highly useful improvement would be to not show trash and junk folders in the unread folder view. In fact, having a way through the UI to customize or create new views for the left side bar would be even better, though less likely to be something that I would expect at this late stage in the release cycle. I’m really enjoying Thunderbird again. Now, if we could just find a way to step up spam filtering a couple notches again, given that Bayesian filtering seems to have been defeated…

About six months ago, I started talking with Atif Khan about a device his company has been working on since 2004. Today I got a chance for the first time to see more details on the deeda, which looks more like an iPhone than anything else I’ve seen. The approach they are taking to application overlays, and really thinking about how social networking services and a device like this would work together is rather compelling and interesting. Deeda is at best a substantial underdog to Apple, and at worst a product that will never make it to market. It will be interesting to see if their platform, which they have described as open source, can survive. I’m also curious to see how many other non-cell phone manufacturers are out there in this space working on similar devices. I of course can’t wait to see Dojo-based web apps running on this emerging class of devices!

A number of random JavaScript musings from recent events, none of which are meaty enough to constitute a blog post.

First off, congrats to Jack Slocum on launching ext and Dean Edwards on unleashing base2. It’s great to see solid work pushing things forward in the JavaScript world.

Coach Wei of Nexaweb wrote an insightful, sad, yet accurate account of Dojo and Ajax performance and bloat. Dojo 0.9 base and core should help a lot with this. Stay tuned for more details on our early summer release of Dojo that will have significant performance improvements and much less code bloat by default. But the other point he raises is that which we have attempted to resolve with a JS Linker. Through significant efforts were made last summer to bring this project to alpha quality, we do not currently have a module owner or anyone actively contributing towards this effort. If this is the sort of project that interests you, please let us know.

Fortify put a paper out about the vulnerabilities of JavaScript Hijacking. For anyone curious of the Dojo response to this, it goes as follows: We worked with Fortify prior to the release of their paper to help them better assess the exposure and to implement remediation in Dojo. However, the most important steps to take with regards to this problem are server-side. Dojo can’t help or hurt your application’s security posture, but we will encourage you to do the right thing. Patches advising users to consider alternate transport methods are scheduled for 0.4.3 and are already available. Also, we’ve offered to coordinate industry consensus on a protocol to make these fixes canonical, but no one has taken us up on it so far.

Finally, we’ll be in London and Paris at the end of May for a Dojo Training Course and the Grails eXchange conference. If anyone wants to meetup, please drop me a line.

We made the annual trek to SxSW this year and had an amazing experience as usual. This year I gave a short talk titled “Web Vector Graphics: Myth or Reality?” and did a Paul Graham style Q&A session at the end with a large number of answers in a very short amount of time. I strongly believe that answering questions is much more valuable than the actual slides I present, so I like to answer as many questions as possible.

On that note, the value of conferences such as SxSW is rarely in the panels themselves. While I really enjoyed Matt Mullenweg’s panel and a panel on bootstrapping, I spend most of my time roaming the halls, chatting with people. Given that we spend most of our time isolated behind our laptops, getting out for a few days and chatting about the latest and greatest with amazingly talented people is why we attend. For example, I finally met John Resig of Mozilla and jQuery, and Andrew Dupont of Prototype, two fellow JavaScript cohorts, and we discussed how dojo.gfx might work as a module that could be plugged into any JavaScript toolkit.
Continue Reading »

sxsw is nearly upon us. It’s that special time of year when the bay area tech community goes on spring break to Austin, TX, for the interactive portion of the festival otherwise known for great music and movies.

This year, I’m doing a solo panel called “Web Vector Graphics: Myth or Reality” where I will talk about browser native vector graphics. The panel starts at 4:05 on Saturday afternoon, and it should be a lot of fun. I usually give talks on more advanced topics, and while I will still do the crowd justice with a fair share of code fragments (there’s never enough code at SXSW), it’s enjoyable to give a talk on something that has been a long time passion and hobby for me: making it possible for the artistically challenged (a.k.a me) to draw a circle.

With Dojo, we’ve recently been discussing splitting up the toolkit into better, easier to digest, less “expensive” code. To which I asked myself, what does expensive mean, for the creators and users of Dojo?

• download size
• number of requests
• CPU utilization
• memory usage
• # of browser/scenario forks
• complexity/time to get started
• error-proneness
• maintainability
• testability
• marketability
• interoperability
• competitiveness
• distribution

With the release of Adium Message style – Renkoo, v2.3, the graphical assets of Renkoo are now available open source under the BSD and AFL license.

This is great news. In addition to providing a great invite service, Renkoo is an example of a new type of company that really gets open source (something we aspire to be with SitePen as well). They don’t just base much of their business model on open source software, but they make significant contributions to the Dojo Toolkit and other efforts. Now, they have taken one of the most popular chat styles, developed originally by Torrey Rice for Renkoo, which they already allowed to be used as a theme in IM clients such as Adium, but now it is completely free to use under the BSD and AFL, meaning that anyone can make use of it in their chat theme or web app!

I’ve been traveling a lot lately, and JetBlue and Southwest recently upgraded their web sites. Given my role in Dojo, I thought it would be interesting to see what JavaScript/Ajax toolkits airline web sites are using (if any), and to see if there is a correlation between the quality of the airline web site and use of a toolkit.

I took a very quick look at the index page source for a number of domestic airlines listed below, and I find JetBlue and Southwest to be among the best. JetBlue currently uses a combination of Prototype-lite, YUI-min, and Lightbox, while Southwest uses YUI. Northwest also uses YUI, though their site is not as pleasant to use. A few of the other airlines listed make use of decent CSS, but have not yet jumped on the Ajax toolkit bandwagon. Unfortunately, none are using Dojo today, but perhaps one or more of the sites below will adopt our toolkit this year!

Domestic airlines surveyed:

Alaska, American, Delta, Frontier, JetBlue, Northwest, United, US Airways

International airlines surveyed:

Aer Lingus, Air Canada, Air France, Air New Zealand, Alitalia, All Nippon, BMI, British Airways, Cathay Pacific, Continental, Emirates, Japan Airlines, KLM, Lufthansa, Ryan Air, SAS, Signapore, Swiss Air, Virgin Atlantic

Because of my use of multiple monitor setups, I have this habit of trying to drag my mouse across screens on different computers. At 3D2, Paul told Torrey about Synergy, which is a lightweight server that lets you do just that. It’s a really really cool, easy to configure way of getting the benefits of the KM portions of a KVM, with the added benefit of being able to copy and paste across computers.